Three-dimensional model testing in the transonic self-streamlining wind tunnel
The wall interference effects present on three-dimensional models during wind tunnel testing are difficult to correct using post-test model data correction methods. Further, at transonic speeds, with the use of ventilated test sections these corrections become complex to apply and inaccurate. The high quality of wind tunnel testing that is required today means that such methods are no longer satisfactory. The flexible walled wind tunnel has in recent years shown its ability to obtain two-dimensional aerofoil data free from the effects of wall boundary restraint. This work at Southampton was aimed at extending the use of the two-dimensional Transonic Self-Streaming Wind Tunnel to the relief of wall interference effects on three-dimensional models. The compromise of using only two-wall single curvature movement avoids the problems that are inherent with the additional complexity of fully three-dimensional adaptive tunnels. A method of assessing the wall-induced interference velocity components from tunnel boundary pressure data, without reference to the model, has been developed and validated against other wall interference assessment methods. The algorithm, suitable for use in adaptive tunnels, is used with a wall movement influence coefficient method of wall contour prediction resulting in the apparent removal of wall interference effects along a streamlining target line. The residual wall interference velocity components calculated to be present after streamlining on two half-wing models are significantly lower than their straight test section values. Providing the model span is not too large in comparison with the breadth of the test section, the spanwise interference velocity component is negligible. A calibrated force-balance wing-body model has been used to demonstrate the first successful streamlining around a three-dimensional model in the Transonic Self-Streamlining Wind Tunnel. The measured model force data obtained with streamlined walls compares favourably with that derived using a standard post-test model data correction method.